Violent but at least Christian: Religious stereotypes and impression formation in a Spanish sample

Fernando Gordillo León fgordilloleon@usal.es

Universidad de Salamanca, España

Amalia Escalona Martínez aescalona@ucjc.edu

Universidad Camilo José Cela, España

Miguel Ángel Pérez Nieto mperez@ucjc.edu

Universidad Camilo José Cela, España

Francisco Pérez-Fernández fperez@ucjc.edu

Facultad HM de Ciencias de la Salud de la UCJC, España

Gabriela Castillo Parra gcastillo@ucjc.edu

Universidad Camilo José Cela, España

José M Arana Martínez arana@usal.es

Universidad de Salamanca, España

Many of the conflicts currently occurring worldwide have their origin in ideological differences between religious groups, which is a recurring theme throughout history (Armstrong, 2014; Cavanaugh, 2009). In recent years this debate has focused on the possibility that "believers" generate higher levels of violence than "non-believers" within conflict situations, with few rigorous empirical studies on this topic (Gingesa et al., 2016). This could be due to many factors, such as the intolerance that most religions maintain and promote in a direct, subtle or ambiguous way, towards non-believers (Bushman et al., 2007), and other religions (Hogemann & Tavares, 2021). Religious people are especially motivated to defend their ideas, regardless of whether these ideas are religious. Religious beliefs are more related to intuition than to analytical and cognitive processes (Pennycook et al., 2016), making it difficult to process contradictory information. This is because a strong tendency is generated to search information that supports our beliefs, ignoring or distorting data that contradict them (Nickerson, 1998; Myers & DeWall, 2015; Peters, 2022).

This intolerance can be explained by stereotypes, which are generalizations about the members of a group that can be positive or negative, the latter being the most frequent and facilitators of prejudice and discrimination. The activation of a stereotype increases the likelihood that the construct it is representing will affect people's subsequent judgments and behaviours. This effect occurs in various contexts and with various consequences (Horcajo et al., 2007). For example, in the employment context, by devaluing workers (Teresa-Morales et al., 2022). In the judicial context, by conditioning the behaviour of judges (Ash et al., 2024). In the sport context, by discouraging people from doing a certain type of sport because it does not conform to a particular stereotype (Rasmussen et al., 2021). It also prevents persons from moving from one social category to another because of the rigidity of social roles (Eagly & Koenig, 2021). Furthermore, it should be considered that the media can contribute to the development and maintenance of stereotypes (Ward & Grower, 2020).

In the religious context, Christians consider the members of their group to be characterized as charitable and compassionate (Simpson & Ríos, 2016). Moreover, activation of the Christian religious stereotype in threat situations generates a kind, selfless, and generous response (Schumann et al., 2014). This makes sense if we consider that religion is part of our self-concept, and therefore has an important influence on our attitudes and behaviours (Ysseldyk et al., 2010). On the other hand, Muslims are negatively stereotyped in the West, as their religious convictions are considered incompatible with Western values (Van Bergen et al., 2017). They are associated with low-paid jobs, violence and women´s oppression (el-Aswad, 2013). In addition, negative beliefs about Muslims are present in the media, strengthening the negative stereotype (Tama & Sulistyaningrum, 2022).

Religious stereotypes affect the relationship between groups. Specifically, there is a tendency to value one's own group (in-group) favourably and the other group (out-group) unfavourably (Hewstone et al., 2002; Jackson & Hunsberger, 1999). In this respect, the ultimate attribution error (Pettigrew, 1979) is the tendency to attribute negative actions of the out-group to internal and stable causes, and those of the in-group to external and punctual causes (Coleman, 2013; Khandelwal et al., 2014). This bias has been tested in priming tasks, where religious stereotype is activated with words to test its effect on attitudes (e.g., Johnson et al., 2012). However, little is known about how the ultimate attribution error might be affecting impression formation.

The way in which others assess your personality has important implications, because personality is a stable construct over time that allows to predict the behaviour of others. First impressions are made in very short time intervals (100 ms approx.) (Willis & Todorov, 2006). Moreover, the way you are perceived by others in a first impression has important consequences in different domains, such as at work, in social, personal, intimate relationships or in the judicial context. The simple presentation of adjectives (prime) next to a person's photograph has an important impact on the way in which that person (target) is subsequently evaluated (Bargh, 2006). Similarly, religious stereotypes can generate evaluative biases. For example, recent research found that Muslims associated with a crime were perceived as cooler and more competent, relative to Christians or atheists who had committed the same crime (Frings et al., 2018). This could have consequences on judicial decisions about the innocence or guilt of the accused. In addition, the interaction between verbal and nonverbal information is also relevant. Specifically, when verbal and nonverbal information is positive, consistent and coding is done separately, the person is considered more stable, friendly and sociable (Gordillo et al., 2017).

The way in which people infer the personality of others has been studied from different approaches. Some works study voice as a predictor of personality (Vinciarelli & Mohammadi, 2014; Pérez-Espinosa et al., 2022), others facial expression (Qin et al., 2016; Gloor, et al., 2022). Other studies use computer software with a multimodal approach (Personality Computing, see Phan & Rauthmann, 2021). Some studies have used priming procedures for the implicit activation of stereotypes. In these investigations, verbal information related to the stereotype is used to activate it (e.g., Kawakami & Dovidio, 2001; Macrae & Cloutier, 2009; Macrae & Martin, 2007). On the other hand, explicit stereotype activation is realized through the manipulation of information given about the characteristics of a group, or through words that allow stereotype activation and that are consciously processed by subjects (e.g., Cardozo et al., 2021; Hess et al., 2004). Therefore, we can infer that words are an instrument that allows the creation of interaction contexts for the study of impression formation (verbal context), as has been shown in the study by Gordillo et al. (2017), where positive and negative words associated with the photograph of a person were selected and presented to a group of subjects who had to evaluate his or her personality. This methodology allows activating religious stereotypes using words such as "Christian-Bible" or "Muslim-Quran", in interaction with different emotional verbal contexts (positive, negative, violent words).

Therefore, and considering the reviewed theoretical context, the objective of this study will be to analyse the influence of religious stereotype activation (Christian and Muslim) and verbal context (neutral and violent words) on personality assessment (openness to experience, agreeableness, responsibility, sociability, and creativity). The following hypothesis is proposed: H1: The Christian sample makes a less negative personality assessment after activation of the violent Christian stereotype than after activation of the violent Muslim stereotype. The acceptance of the hypothesis would reflect the effect of the final attribution error on personality assessment.

The aim of this research was to analyse the effect of religious stereotype activation (Christian and Muslim) in interaction with verbal context (neutral and violent) on personality appraisal (emotional stability, agreeableness, responsibility, sociability and creativity). The results allow us to accept the hypothesis because the personality assessment was more negative after the activation of the violent Muslim stereotype than after the activation of the violent Christian stereotype. These results can be explained by the ultimate attribution error, which states that ingroup members tend to estimate that the causes of the negative behaviours (violent) of the outgroup are explained by internal causes such as personality, and not so much by external causes (e.g., a given situation). In this case, the violence perceived after the activation of the violent Muslim stereotype generated a more negative evaluation of the personality (vs. violent Christian stereotype), which can be interpreted as follows: the sample (Christian) considers that the cause of the violence reflected by the words is internal, and consequently they make a more negative evaluation of the personality of a component of the Muslim out-group.

When we make an internal attribution (personality) about violent behaviour of a person, without considering other possible causes (e.g., stressful situation), we are biasing the processing of information. This has a negative effect because we assume that this person will be violent in many other situations, when this does not have to be true. This may not matter in some situations, but when religious groups are involved, it implies that the negative stereotype of the out-group is being reinforced by assuming that any violent behaviour by members of that group is explained by internal causes. Different studies have reported this bias in attributing causation of socially undesirable behaviour among religious groups (Taylor and Jaggi, 1974; Islam and Hewstone, 1993; Hewstone and Ward, 1985; Khan and Liu, 2008); and different metanalyses have consistently supported the intergroup bias (ultimate attribution error), also reporting moderators and mediators, such as culture, with evidence that Asians are less susceptible to this bias. In addition, empathy, promoted through intergroup contact and training can reduce this bias. Even simply knowing that a group member has friends in the other group improves attitudes toward the other group (Beelmann & Heineman, 2014; Lemmer & Wagner, 2015; Pettigrew & Tropp, 2006; Zhou et al., 2019).

This research has focused on religion as the origin of the differences between groups, but this perspective is limited if we do not consider that violence can have its origin in many other aspects and can be explained from different theoretical approaches. Many theories try to explain the origin of wars as something inherent to human beings (Jiménez-Bautista, 2012): 1) within the Christian religion humans would be stigmatized by "original sin", thus justifying all violent and negative actions; 2) within individualism violence would have a place as an instrument to achieve one's own benefit, without the need to assume the general rules; 3) within social Darwinism violence would play a role in evolution, under the premise that natural selection benefits the strongest individuals. This is not a simple subject and requires a transversal perspective in accordance with the complexity of human behaviour, from a determining biological basis to the modulating social and cultural context.

Some strategies could be applied to minimise biases in information processing caused by stereotypes. This is especially important in contexts such as the judiciary or health care. (Van Ryn, 2016): 1) Put yourself in someone’s shoes; 2) Build partnerships with others; 3) Take care of yourself: 4) protect your mental resources; 5) Be positive; 6) Counteract negative stereotypes by exposing yourself to positive images. One of the limitations of this work is the number of participants, although the sample is sufficiently heterogeneous in age and gender to be representative. On the other hand, we only presented the photograph of one person in phase 2 and 3. This was necessary because the experiment was performed online, and we needed a limited number of trials to avoid demotivation and fatigue of the participants. In addition, the word "Christian" does not discriminate between Catholics and Protestants, something that would have to be considered in the interpretation of the results, because these groups could consider each other as members of a different social group, regardless of whether they both integrate within Christianity.

Future research should investigate the effect that religious stereotypes, in interaction with the positive verbal context, have on personality assessment. Surely the positive context plays a determining role in this process, as a counterpoint to the results obtained in the present investigation, facilitating integration between the different religious groups living together in the same territory. In conclusion, it can be said that personality assessment is a variable sensitive to the attributional processes present in the relationships between groups, especially the dimensions of emotional stability, agreeableness and sociability, and that the differences found, since they refer to internal, stable and difficult to modify characteristics, would increase the sense of frustration between groups with respect to the possibility of integration and/or reconciliation of positions. This, ultimately, could reinforce the negative stereotypes of religious groups, making it difficult to resolve conflicts.